parse-html-table/dist/index.mjs

A lightweight library for parsing HTML tables into structured data using TypeScript.

import cheerio from 'cheerio';

function createTableParser() {
  return (html) => {
    const $ = cheerio.load(html);
    let headers = [];
    const parseHeaders = () => {
      headers = $("th").map((_, el) => {
        const headerText = $(el).text().trim();
        const words = headerText.split(/\s+/);
        const camelCaseText = words.map((word, index) => {
          if (index === 0) {
            return word.toLowerCase();
          } else {
            return word.charAt(0).toUpperCase() + word.slice(1).toLowerCase();
          }
        }).join("");
        return camelCaseText;
      }).get();
    };
    const parseRows = () => {
      return $("tbody tr").map((_, row) => {
        const rowData = {};
        const $cells = $(row).find("td");
        let isEmptyRow = true;
        $cells.each((index, cell) => {
          const cellText = $(cell).text().trim();
          if (cellText !== "") {
            rowData[headers[index]] = cellText;
            isEmptyRow = false;
          }
        });
        return isEmptyRow ? null : rowData;
      }).get().filter((rowData) => rowData !== null);
    };
    const parseTable = () => {
      parseHeaders();
      const rows = parseRows();
      return {
        headers,
        rows
      };
    };
    return parseTable;
  };
}

export { createTableParser };